Research

JAMA Psychiatry | Original Investigation

Effect of Psychiatric Advance Directives Facilitated by Peer Workers

on Compulsory Admission Among People With Mental Illness
A Randomized Clinical Trial
Aurélie Tinland, MD, PhD; Sandrine Loubière, PhD; Frederic Mougeot, PhD; Emmanuelle Jouet, PhD;
Magali Pontier, MD; Karine Baumstarck, MD, PhD; Anderson Loundou, PhD; Nicolas Franck, MD, PhD;
Christophe Lançon, MD, PhD; Pascal Auquier, MD, PhD; for the DAiP Group

Supplemental content
IMPORTANCE Reducing the use of coercion in mental health care is crucial from a human
rights and public health perspective. Psychiatric advance directives (PADs) are promising
tools that may reduce compulsory admissions. Assessments of PADs have included
facilitation by health care agents but not facilitation by peer workers.

OBJECTIVE To determine the efficacy of PADs facilitated by peer workers (PW-PAD) in people
with mental disorders.

DESIGN, SETTING, AND PARTICIPANTS A multicenter randomized clinical trial was conducted in
7 French mental health facilities. Adults with a DSM-5 diagnosis of schizophrenia, bipolar I
disorder, or schizoaffective disorder who had a compulsory admission in the past 12 months
and the capacity to consent were enrolled between January 2019 and June 2020 and
followed up for 12 months.

INTERVENTIONS The PW-PAD group was invited to fill out a PAD form and meet a peer worker
who was trained to assist in completing and sharing the form with relatives and psychiatrists.

MAIN OUTCOMES AND MEASURES The primary outcome was the rate of compulsory admission
at 12 months after randomization. The overall psychiatric admission rate, therapeutic alliance,
quality of life, mental health symptoms, empowerment, and recovery outcomes were also
investigated.

RESULTS Among 394 allocated participants (median age, 39 years; 39.3% female; 45% with
schizophrenia, 36% bipolar I disorder, and 19% schizoaffective disorder), 196 were assigned
to the PW-PAD group and 198 to the control group. In the PW-PAD group, 27.0% had
compulsory admissions compared with 39.9% in the control group (risk difference, −0.13;
95% CI, −0.22 to −0.04; P = .007). No significant differences were found in the rate of overall
admissions, therapeutic alliance score, and quality of life. Participants in the PW-PAD group
exhibited fewer symptoms (effect size, −0.20; 95% CI, −0.40 to 0.00), greater
empowerment (effect size, 0.30; 95% CI, 0.10 to 0.50), and a higher recovery score (effect
size, 0.44; 95% CI, 0.24 to 0.65), compared with those in the control group.

CONCLUSIONS AND RELEVANCE Peer worker–facilitated PADs are effective in decreasing

compulsory hospital admissions and increasing some mental health outcomes (self-perceived
symptoms, empowerment, and recovery). Involving peer workers in the completion of PADs
supports the current shift of mental health care from substitute decision-making to
supported decision-making.

TRIAL REGISTRATION ClinicalTrials.gov Identifier: NCT03630822

Author Affiliations: Author

affiliations are listed at the end of this
article.
Group Information: Additional
members of the DAiP Group appear
in Supplement 3.
Corresponding Author: Aurélie
Tinland, MD, PhD, EA3279–CEReSS,
Faculté de Médecine Timone, 27
Boulevard Jean Moulin, F-13005
JAMA Psychiatry. 2022;79(8):752-759. doi:10.1001/jamapsychiatry.2022.1627 Marseille Cedex 5, France (aurelie.
Published online June 6, 2022. tinland@gmail.com).

752 (Reprinted) jamapsychiatry.com

Downloaded From: https://jamanetwork.com/ on 06/12/2023

 Effect of Peer Worker–Facilitated Psychiatric Advance Directives in People With Mental Illness
R
espect for patient autonomy is such a strong pillar of
health care that involuntary treatment should be
unusual. Despite this ethical and clinical principle, com-
pulsory psychiatric admissions are far too common in coun-
tries of all income levels.1-3
In the last 20 years, several randomized clinical trials (RCTs)
have assessed the effectiveness of interventions in reducing
compulsory psychiatric admissions, and systematic reviews
showed that the most effective were psychiatric advance
statements.4 Psychiatric advance statements are written docu-
ments that allow adults who have mental illness to state their
will and preferences in advance so their choices can be applied
if further mental health crises impair their decision-making ca-
pacity. A meta-analysis of the 5 most robust RCTs on this topic
was published in 2019.5 Molyneaux et al5 suggested that (1) psy-
chiatric advance statements reduced the risk of compulsory ad-
mission among individuals with mental disorders by 25% com-
pared with usual care; (2) RCTs did not provide firm conclusions
on other criteria such as therapeutic alliance or psychiatric out-
comes; and (3) similar and higher effectiveness was found in
pooled studies that addressed interventions with crisis plan-
ning and facilitation by health care professionals.
Other research highlights the importance of facilitation in
completing psychiatric advance statements,6 but having health
care professionals serve as facilitators is not an obvious choice.
Indeed, a climate of coercion in psychiatry has been de-
scribed, with informal coercion and use of power beyond for-
mal coercion, that is, involuntary admissions and compul-
sory treatment.7,8 The relationship between patients and
health care professionals is permeated by this coercive cli-
mate more than clinicians realize.9 Because autonomy and self-
determination are the main tenets of psychiatric advance
statements,10 this system can be improved if facilitation is pro-
vided by other types of professionals who are less likely to ex-
ert undue influence. To date, RCTs have assessed facilitation
by researchers11 and patient advocates,12 but their results are
not as significant as those where health care professionals were
the facilitators.
In France, advance directives were created by law in 2005
and primarily used in end-of-life health care. Psychiatric ad-
vance directives (PADs) have been used without a formal legal
or practical framework. We hypothesized that PADs could be
implemented with facilitation by peer workers, ie, people with
personal experiences of mental distress and psychiatric ser-
vices who are employed and trained to support others.13 Inter-
est in the facilitation of PADs by peer workers is increasing, and
a comparison between peer workers and health care agents
found no differences in PAD completion rate and quality.14 Re-
cent studies showed that PADs facilitated by peer workers were
more prescriptive than those facilitated by nonpeer clinicians
and had high feasibility and consistency as rated by experts.15
To our knowledge, no study has been conducted of the ef-
fect on clinical outcomes of peer worker–facilitated PADs (PW-
PADs). The current study addresses this evidence gap to de-
termine whether PW-PADs for people with severe mental illness
reduce compulsory admissions and provide significant ben-
efits in terms of therapeutic alliance, quality of life, mental ill-
ness symptoms, empowerment, and recovery.
jamapsychiatry.com
Downloaded From: https://jamanetwork.com/ on 06/12/2023
Original Investigation Research
Key Points
Question Do psychiatric advance directives facilitated by peer
workers (PW-PAD) reduce coercion and improve clinical outcomes
among people with schizophrenia, bipolar I disorder, or
schizoaffective disorder?
Findings In this randomized clinical trial, 394 adults with
schizophrenia, bipolar I disorder, or schizoaffective disorder with a
previous compulsory hospitalization were randomized into 2
groups (ratio 1:1). Participants in the PW-PAD group experienced
significantly fewer compulsory admissions than those in the
control group.
Meaning These findings support the use of peer
worker–facilitated psychiatric advance directives to prevent
compulsory rehospitalization in people with severe mental illness.
Methods
Ethics
The trial was submitted and approved June 6, 2018, by the
French ethics committee Sud-Ouest et Outre-mer 4 (2018-
A00146-49). The study was conducted in compliance with the
Declaration of Helsinki, sixth revision; Good Clinical Practice
guidelines; and local regulatory requirements. The partici-
pants provided both oral and written consent before their en-
rollment and allocation to the study groups.
Trial Design
The study, which has the English name description Peer Worker–
Facilitated Psychiatric Advance Directive Study (French acro-
nym DAiP), was a multicenter nonblinded RCT conducted in 7
mental health facilities (aka centers) of 3 cities (aka sites) in
France (Lyon, Paris, and Marseille) between January 2019 and
June 2021. Participants were referred by their treating psychia-
trists in mental health institutions. At the time of study inclu-
sion, most participants were discharged from the hospital, but
some were still hospitalized. Psychiatrists checked the eligibil-
ity criteria and referred eligible participants to research assis-
tants. Research assistants and psychiatrists reviewed the pa-
tient information and validated the inclusion and exclusion
criteria. Research assistants met participants from both groups
at a location of their choice for face-to-face interviews at the time
of inclusion and at 6 and 12 months. The 12-month follow-up
timeline started directly after randomization. The recruitment
period was planned for 12 months but extended by 6 months
because of the COVID-19 pandemic. The study was stopped as
originally planned 12 months after the last recruitment. Full de-
tails are available in the published protocol (Supplement 2).16
The study is registered at ClinicalTrials.gov (NCT03630822).
This article follows the Consolidated Standards of Reporting Trials
(CONSORT) reporting guidelines.
Population and Randomization
Eligible participants were older than 18 years; were involun-
tarily admitted to the hospital within the past 12 months; had
(Reprinted) JAMA Psychiatry August 2022 Volume 79, Number 8 753
 Research Original Investigation Effect of Peer Worker–Facilitated Psychiatric Advance Directives in People With Mental Illness
a diagnosis of schizophrenia, bipolar I disorder, or schizoaf-
fective disorder according to DSM-5 criteria17; had decision-
making capacity assessed by a psychiatrist according to the
MacArthur Competence Assessment Tool for Clinical
Research18; were covered by French government health insur-
ance; and understood French. The exclusion criteria in-
cluded being considered unable to provide informed consent
and being under tutorship (the more restrictive of 2 levels of
guardianship in France).
Immediately after signing the consent form, participants
were randomly assigned using a web-based system at a 1:1 ra-
tio. The randomization list used a permuted block design and
was stratified by the center. Research assistants, treating cli-
nicians, and participants were aware of the assigned random-
ization group.
Intervention Group (PW-PAD)
After randomization, all PW-PAD participants received the PAD
document from research assistants. The PAD documents in-
cluded future treatment and support preferences, early signs
of relapse, and coping strategies. The research assistant pro-
posed organizing the meeting with the peer worker and dis-
tributed the contact details. Depending on their preferences,
the PW-PAD participants could:
• Meet a peer worker in a place of their choice.
• Be supported by this peer worker in drafting the PAD docu-
ment with as many meetings as necessary. At this stage, the
peer workers encouraged the sharing of PADs.
• Be supported by the peer worker during the sharing of PADs
with the health care agent and the psychiatrist.
The PADs were completed and signed in paper format.
They were stored by the health care agent or the psychiatrist,
depending on the choice of the participant, and uploaded to
electronic medical records if available and requested. In case
of crisis, the existence of a PAD was reported by the patient,
their companions, or informed caregivers.
Control Group
People assigned to the control group were followed up as usual.
They received comprehensive information about the PAD con-
cept during the inclusion step and were free to complete a PAD.
They were not introduced to a peer worker from the study.
Outcomes
We collected data from the computerized patient administra-
tive system and through participant interviews, which were
scheduled every 6 months. The primary outcome was the rate
of compulsory admissions to a psychiatric hospital at 12 months
of follow-up, calculated as the number of participants with at
least 1 compulsory admission divided by the number of par-
ticipants.
Secondar y outcomes were c are-related, patient-
reported, and mental health outcomes. The care-related out-
comes included overall hospital admission rate (including vol-
untary and involuntary admissions), total number of
admissions per patient (including voluntary and involuntary
admissions), and rate of noncompulsory admissions per pa-
tient (ie, the proportion of total admissions per patient that was
754 JAMA Psychiatry August 2022 Volume 79, Number 8 (Reprinted)
Downloaded From: https://jamanetwork.com/ on 06/12/2023
noncompulsory). Therapeutic alliance was assessed using the
4-point ordinal Alliance Scale.19 Higher scores indicate higher
therapeutic alliance.
Among the patient-reported outcomes, quality of life was
assessed using the Schizophrenia Quality-of-Life scale.20
Dimension and index scores range from 0, which indicates the
lowest quality of life, to 100, which indicates the highest qual-
ity of life. Health status was assessed using the EuroQol scale
(5 dimensions and 3 Likert).21 The index score ranges from 0,
which indicates the worst health, to 1, which indicates the best
health.
Mental-health outcomes included symptomatology as-
sessed with the self-reported modified Colorado Symptom
Index.22 Higher scores indicate a greater likelihood of mental
health problems. Empowerment was assessed using the
Empowerment Scale.23 The index scores are 0 to 100, where
higher scores correspond to higher empowerment. Recovery
was assessed using the Recovery Assessment Scale. A higher
score indicates better recovery.
Other individual data collected at baseline included so-
ciodemographic information and clinical data. Gender, age,
education level, nationality, social benefits, wages, employ-
ment status, and housing conditions were collected. Depriva-
tion was assessed using the EPICES score (English descrip-
tion of score name: Evaluation of Deprivation and Inequalities
in Health Examination Centers).24 Clinical data assessed by the
psychiatrist included somatic and addictive comorbidity and
overall condition using the Clinical Global Impression scale,
scored from 1 (healthy, not ill) to 7 (severely ill). Secondary out-
comes are described in more detail in the eMethods in Supple-
ment 1.
Sample Size
The sample size was calculated to detect a reduction of 30%
in the rate of compulsory admissions to psychiatric hospitals
during the follow-up period of 12 months between the 2
groups12,25,26 with a reference point of 42.6%.27 To obtain a sig-
nificance level of 2.5% and power of 80% with equal alloca-
tion to 2 groups, each group of the trial required 182 people.
To allow for a potential 10% of people being lost to follow-up,
the planned sample was 200 per group, for a total of 400.
Statistical Analysis
The intention-to-treat analysis included all randomized par-
ticipants. Data on compulsory admissions were obtained for
all participants; no imputation was performed for the pri-
mary outcome. For secondary outcomes, missing data due to
withdrawal, loss to follow-up, or nonresponse to specific items
were 10% to 37.2% at 12 months. Missing data were ad-
dressed using multiple imputations,28 which creates mul-
tiple “complete” data sets with predictions for each missing
value. This procedure takes into account uncertainty and yields
accurate standard errors.29 Fifty imputed data sets were imple-
mented using MICE by chained equations and mitools R pack-
ages. The multiple imputation approach was compared with
existing methods for handing missing data; for complete cases,
imputed data were compared with the mean or the last obser-
vation carried forward.
jamapsychiatry.com
 Effect of Peer Worker–Facilitated Psychiatric Advance Directives in People With Mental Illness Original Investigation Research

Data analysis was conducted in 3 steps. First, we per-

Figure. CONSORT Flow Diagram for the DAiP Trial
formed a collinearity test on potential confounding factors (Peer Worker–Facilitated Psychiatric Advance Directive Study)
based on unbalanced baseline characteristics. No collinearity
was observed; the variance inflation factor ranged from 1.041 473 Assessed for eligibility from January 2019 to June 2020
to 1.112. Second, the proportion of patients with compulsory
admission was compared between groups using generalized 72 Excluded by research assistants
estimating equations (GENLIN function), applying a bino- 48 Declined to participate
(not interested, worried
mial distribution with a link logit and adjusting for unbal- that drafting PADs may
anced baseline covariates (age, diagnosis, and Clinical Global lead to a crisis)
14 Had no contact after reflection
Impression score), as well as site and site × group interaction. 10 Did not meet inclusion criteria
No group × covariate interactions were kept in the model (they
were nonsignificant). Further, the subsequent model applied
401 Randomized
a logistic regression that provided a high goodness of fit. Ad-
justed odds ratios and risk differences with 95% CI were
calculated. 200 Randomized to intervention group 201 Randomized to control group

For secondary outcomes, between-group differences

4 Excluded by data board 3 Excluded by data board
were estimated using generalized estimating equations
3 Withdrew 2 Withdrew
(GENLIN function), applying a normal distribution with a 1 Eligibility criteria 1 Eligibility criteria
not verified not verified
link identity for score variables or Poisson distribution with a
link log for count variables. The β coefficient and effect sizes
196 Individuals in intervention group 198 Individuals in intervention group
(Cohen d) with 95% CI were calculated. Statistical analysis
was performed using SPSS 12 for Windows and RStudio ver-
196 With primary outcome data at 198 With primary outcome data at
sion 3.2.1. 6-mo follow-up 6-mo follow-up
118 Completed the interview 127 Completed the interview
78 Discontinued 71 Discontinued
78 Discontinued and did not 69 Did not attend 6-mo
attend 6-mo follow-up follow-up
Results 1 Died
1 Withdrew
Participants
As depicted in the Figure, 401 patients were randomized in 196 With primary outcome data at 198 With primary outcome data at
total, of whom 7 (1.7%) were excluded from the study by the 12-mo follow-up 12-mo follow-up
127 Completed the interview 139 Completed the interview
data board (4 patients allocated to the PW-PAD group and 3 69 Discontinued 60 Discontinued
patients allocated to the control group). Two were excluded 68 Did not attend 12-mo 56 Did not attend 12-mo
follow-up follow-up
because of noneligible inclusion criteria and 4 because they 1 Withdrew 1 Died
withdrew before any data collection. Of the 394 patients 2 Withdrew
included in the study, 196 were assigned to the intervention
group, and 198 were assigned to the control group (Figure). 196 Analyzed in ITT analysis 198 Analyzed in ITT analysis

Interviews at the 12-month follow-up were completed for

127 (65%) in the PW-PAD group and 139 (70%) in the control
group.
Baseline characteristics were similar between groups, ex-
cept for age and severity (Table 1). Most patients were male
(239, 60.7%) and had completed postsecondary school edu-
cation (261, 66.4%). Of the 3 diagnoses assessed, 139 partici-
pants (36%) had bipolar I disorder, 178 (45%) schizophrenia,
and 76 (19%) schizoaffective disorder. The median (IQR) age
of the sample was 39 (29-48) years, and participants in the in-
tervention group were younger (mean [SD], 37.4 [11.7] years
vs 41.0 [12.7] years; P = .003). Baseline characteristics per site
are provided in eTable 1 in Supplement 1. Characteristics at
baseline were compared between complete and incomplete
cases at the 12-month follow-up; significant differences were
found in EPICES score, comorbidities, and previous hospital
admissions (eTable 2 in Supplement 1).
Sixty-three participants in the intervention group (31.2%)
and 67 in the control group (33.8%) were included during a hos-
pitalization, with no significant differences in median (IQR)
number of inpatient days between inclusion and discharge (34
[7-76] days vs 36 [11-67] days; P = .73).
ITT indicates intention to treat; PADs, psychiatric advance directives.
Completion of PADs
In the PW-PAD group, 107 participants completed a PAD docu-
ment (54.6%) compared with 14 (7.1%) in the control group
(P < .001). Among those, 81 met facilitators (75.7%), and 29
used PADs during a crisis in the 12-month follow-up period
(27.1%) (Table 2).
Primary Outcome
Table 3 presents the number and percentage of patients who
had compulsory admissions to the hospital during the 12-
month follow-up. The rate of compulsory admission was sig-
nificantly lower in the PW-PAD group than in the control group:
27.0% (53 patients) vs 39.9% (79 patients), respectively (ad-
justed odds ratio, 0.58; 95% CI, 0.37 to 0.92; risk difference,
−0.13; 95% CI, −0.22 to −0.04; P = .007).
Secondary Outcomes
Secondary outcomes are presented in Table 3. Participants in
the PW-PAD group exhibited lower symptoms as measured by

jamapsychiatry.com (Reprinted) JAMA Psychiatry August 2022 Volume 79, Number 8 755

Downloaded From: https://jamanetwork.com/ on 06/12/2023

 Research Original Investigation Effect of Peer Worker–Facilitated Psychiatric Advance Directives in People With Mental Illness

Table 1. Sociodemographic and Clinical Characteristics of Participants

(N = 394) Discussion
Group, No. (%)
Among 394 participants living with schizophrenia, bipolar I dis-
PW-PAD Control
Characteristic (n = 196) (n = 198) order, or schizoaffective disorder who had compulsory hos-
Men 127 (64.8) 112 (56.6) pital admissions during the past year, use of PADs facilitated
Women 69 (35.2) 86 (43.4) by peer workers was associated with a significant decrease in
Age, y compulsory admissions and an increase in mental health out-
Mean (SD) 37.4 (11.7) 41.0 (12.7)
comes (self-perceived symptoms, empowerment, and recov-
ery) at 12 months. In the PW-PAD group, 54.6% of partici-
Median (IQR) 36 (28-44) 40 (31-49)
pants completed PADs (vs 7.1% in control group), among whom
French nationality 184 (93.9) 180 (91.8)
75.7% used the support of peer workers.
Education
This study is the first to our knowledge to show that PADs
Less than HS 57 (29.2) 75 (37.9)
facilitated by peer workers are effective in reducing compul-
Completed HS or postsecondary 138 (70.8) 123 (62.1)
school sory admissions. With a decrease of 32% of compulsory ad-
Marital status missions, these results exceed the 25% pooled estimates from
Single 132 (67.3) 128 (64.6) the meta-analysis. As in all other comparable studies, we found
little effect on overall admissions, which supported that
Married/partnered 38 (19.4) 35 (17.7)
PADs might reduce compulsory admissions by making par-
Divorced/separated/widowed 26 (13.3) 35 (17.7)
ticipants more willing to consider a voluntary admission when
Employed 33 (18.8) 37 (19.9)
a crisis occurs instead of preventing hospital admissions.5
EPICES score
This result is important because minimizing compulsory
Mean (SD) 40.6 (19.9) 42.8 (20.9)
admissions reduces its many widely described negative
Median (IQR) 40.8 (24-57) 44.6 (26-59)
consequences.30-32 The qualitative research conducted in par-
DSM-5 diagnosis allel with this study will provide insight into the mechanisms
Schizophrenia 86 (44.1) 92 (46.5) and drivers of the intervention’s effectiveness.
Bipolar I disorder 66 (33.8) 73 (36.9) We found high rates of involuntary hospitalizations in both
Schizoaffective disorder 43 (22.1) 33 (16.7) groups, which are consistent with published French data.27
Alcohol dependence 6 (3.4) 6 (3.5) France ranked above the international median calculated in
Substance dependence 22 (12.6) 24 (13.6) 2017 among 22 countries, with an annual rate of 140.0 invol-
≥1 Somatic comorbidity 120 (61.2) 137 (69.2) untary hospitalizations per 100 000 people compared with a
CGI score median rate of 106.4 (IQR, 58.5-150.9).33
Mean (SD) 4.1 (1.2) 4.3 (1.1) No previous RCTs on the efficacy of advance statements re-
Median (IQR) 4.0 (3-5) 4.0 (4-5) ported results of mental health outcomes, except Papageorgiou
No. of admissions in previous 1 y, 1.5 (0.9) 1.4 (0.8)
et al11 and Lay et al,34 who did not find any differences in psychi-
mean (SD) atric symptoms and psychiatric functioning at 12 months.
Patients with admissions In our study, PW-PAD was associated with improvement in
in previous 1 y, No. (%)
mental health outcomes, with improvements in symptoms,
1 Admission 132 (67.3) 148 (75.5)
empowerment, and recovery. Research on recovery-oriented
2 Admissions 45 (23.0) 37 (18.9)
services distinguishes between health-related outcome mea-
≥3 Admissions 17 (8.7) 11 (5.6)
sures (such as symptoms) and recovery-oriented outcome mea-
Abbreviations: CGI, Clinical Global Impression scale; EPICES, Evaluation of sures (such as self-assessment of recovery, empowerment, or
Deprivation and Inequalities in Health Examination Centers (English quality of life), which capture the efficacy of peer support more
description); HS, high school; PW-PAD, peer worker–facilitated psychiatric
advance directive.
accurately.35 Research on peer support demonstrates substan-
tial heterogeneity in terms of quality36 and encompasses many
activities,37 but reviews have shown that the involvement of
peers in various services is associated with mixed and lim-
modified Colorado Symptom Index score (effect size, −0.20; ited improvements in recovery-oriented outcomes.37-39 Be-
95% CI, −0.40 to 0.00; P = .03), greater empowerment through cause the PW-PAD intervention had the greatest effect on sev-
Empowerment Scale scores (0.30; 95% CI, 0.10 to 0.50; eral of these indicators, we hypothesize that peer-worker
P = .003), and higher recovery through Recovery Assessment involvement plays a role in these results. Further studies should
Scale scores (0.44; 95% CI, 0.24 to 0.65; P < .001), compared directly compare facilitation by health care professionals and
with the control group. We found no statistically significant facilitation by peer workers using a measurement of per-
differences between groups for overall admission rate, thera- ceived coercion.
peutic alliance, and quality-of-life measures. Furthermore, the high rates of completion in the PW-
Sensitivity analyses showed that the use of multiple im- PAD group show the importance of encouragement (explana-
putations rather than other methods of handling missing data tion, distribution of the document) and facilitation, which
had little effect on the results (eTable 3 in Supplement 1). reinforces previous research.5,6,14

756 JAMA Psychiatry August 2022 Volume 79, Number 8 (Reprinted) jamapsychiatry.com

Downloaded From: https://jamanetwork.com/ on 06/12/2023

 Effect of Peer Worker–Facilitated Psychiatric Advance Directives in People With Mental Illness Original Investigation Research

Table 2. Outcomes at 12 Months Regarding Psychiatric Advance Directives for All Participants (N = 394)
PW-PAD group Control group Total
Outcome (n = 196) (n = 198) (N = 394)
Completion of PAD, No. (%) 107 (54.6) 14 (7.1) 121 (30.7)
Written with peer-worker support, No. (%) 81 (41.3) 4 (2.0) 85 (21.6)
% Among those who completed PAD in intervention 75.7 NA NA
group (n = 107)
Use of PAD during subsequent crisis, No. (%) 29 (14.8) 5 (2.5) 34 (8.6)
% Among those who completed PAD in intervention 27.1 NA NA
group (n = 107)
Abbreviations: NA, not applicable;
Compliance with PAD, No. (%) 22 (11.2) 5 (2.5) 27 (6.8)
PAD, psychiatric advance directive;
% Among those who completed PAD in intervention 20.6 NA NA PW-PAD, peer worker–facilitated
group (n = 107) psychiatric advance directive.

Table 3. Compulsory Admissions, Overall Psychiatric Admissions, and Secondary Outcomes: Regression-Model Results at 12 Months
Between Participants in the PW-PAD Group and Control Group

No. (%) or mean (SD) Logistic regression or GLM models Effect size
PW-PAD group Control group β Coefficient aOR Risk difference Cohen d
(n = 196) (n = 198) (95% CI) (95% CI)a (95% CI)b (95% CI)c
Primary outcome
Patients with ≥1 psychiatric compulsory 53 (27.00) 79 (39.90) −0.57 0.58 −0.13 NA
admission, No. (%) (−1.01 to −0.08)d (0.37 to 0.92)d (−0.22 to −0.04)d
Secondary outcomes
Patients with ≥1 psychiatric admission, 70 (35.70) 79 (39.90) 0.15 1.16 −0.04 NA
No. (%) (−0.29 to 0.59) (0.75 to 1.80) (−0.13 to 0.54)
No. psychiatric admissions per patient, 0.93 (2.19) 1.09 (2.02) −0.16 NA NA −0.08
mean (SD) (−0.64 to 0.25) (−0.30 to 0.12)
Rate of noncompulsory admissions 0.56 (0.45) 0.45 (0.45) 0.21 NA NA 0.47
per patient, mean (SD) (−0.07 to 0.50) (−0.17 to 1.11)
Score for each scale, mean (SD)
4-PAS 35.62 (10.88) 31.56 (9.34) 1.83 NA NA 0.19
(−0.35 to 4.13) (−0.03 to 0.41)
S-QOL 62.39 (21.64) 57.62 (18.73) 3.77 NA NA 0.18
(−0.39 to 7.94) (−0.02 to 0.39)
EQ5D-3L 0.82 (0.27) 0.76 (0.32) 0.03 NA NA 0.17
(−0.01 to 0.06) (−1.08 to 14.0)
MCSI 11.49 (11.91) 13.87 (10.99) −2.38 NA NA −0.20
(−4.59 to −0.18)d (−0.40 to 0.00)d
ES 16.80 (26.32) 10.20 (16.04) 6.05 NA NA 0.30
(1.56 to 10.53)d (0.10 to 0.50)d
RAS 72.60 (14.13) 65.55 (13.92) 6.26 NA NA 0.44
(3.29 to 9.23)d (0.24 to 0.65)d
Abbreviations: 4-PAS, 4-point ordinal Alliance Scale; aOR, adjusted odds ratio; the risk difference (with 95% CI computed).
EQ5D-3L, EuroQol scale at 5 dimensions and 3 Likert; ES, Empowerment Scale; c
Generalized linear models (using either a binomial distribution with a link logit,
GLM, generalized linear model; MCSI, modified Colorado Symptom Index; a negative normal distribution with a link log, or a Poisson distribution with a
NA, not applicable; PW-PAD, peer worker–facilitated psychiatric advance link log for count variables) adjusting for age, diagnosis, Clinical Global
directive; RAS, Recovery Assessment Scale; S-QOL, Schizophrenia Quality of Impression score, and site and group × covariate interaction. A random effect
Life. (site) was fitted in the model with an exchangeable covariance matrix. No
a
Logistic regression adjusting for age, diagnosis, Clinical Global Impression interactions were kept because none achieved statistical significance. The
score, and site (site × group interaction was tested in GLM models and did not effect sizes were estimated from the mean difference divided by the pooled
achieve statistical significance). High goodness of fit: Akaike information SD (using Cohen d formula) and based on the imputed-analysis set.
criterion, 436.6, compared with quasi-likelihood independence model criteria d
Statistically significant difference from the group variable (PW-PAD vs control
in GLM, 496.3. Adjusted odds ratios were reported for group variable. groups).
b
Effect sizes were estimated from the difference in proportions and referred to

Strengths and Limitations
Our study has 2 major strengths. First, the nature of this re-
search is highly participatory because it involved patients at
all levels from the beginning.16 Second, it was deployed through
more than 40 psychiatrists from all backgrounds with cur-
rent practices. Because peer workers were independently re-
cruited by research teams, participating units were not only
recovery-oriented but reflected a variety of services. Thus, the
study suggests that PW-PADs can be easily implemented and
our results are generalizable to other services.
This trial had several limitations. First, the follow-up was
complicated by the COVID-19 health crisis, and the loss of 31%
of participants at the 12-month follow-up and resulting de-
crease in power for secondary outcomes are important limi-
tations. Fortunately, our primary outcome was based on ad-
ministrative data and consequently not affected. Second, the
recruitment was unequal among centers, and we did not have
the power to make comparisons across the 7 different cen-
ters. Third, the profession of peer worker is relatively new in
France, and the PAD was a new tool. The study has led to the
development of specific training for peer workers and teams,
which should help improve the results of such interventions.
Further, we had notably restrictive criteria, and the findings
may not be generalizable to other psychiatric populations.

jamapsychiatry.com (Reprinted) JAMA Psychiatry August 2022 Volume 79, Number 8 757

Downloaded From: https://jamanetwork.com/ on 06/12/2023

 Research Original Investigation
Conclusions
Among people living with schizophrenia, bipolar I disorder, or
schizoaffective disorder, the use of PW-PADs was associated with
a significant decrease in compulsory admissions and a signifi-
ARTICLE INFORMATION
Accepted for Publication: May 5, 2022.
Published Online: June 6, 2022.
doi:10.1001/jamapsychiatry.2022.1627
Open Access: This is an open access article
distributed under the terms of the CC-BY License.
© 2022 Tinland A et al. JAMA Psychiatry.
Author Affiliations: CEReSS–Health Service
Research and Quality of Life Center (UR 3279),
Aix-Marseille University, School of Medicine–La
Timone Medical Campus, Marseille, France (Tinland,
Loubière, Baumstarck, Loundou, Lançon, Auquier);
Department of Psychiatry, Assistance Publique–
Hôpitaux de Marseille, Marseille, France (Tinland,
Pontier, Lançon); Support Unit for Clinical Research
and Economic Evaluation, Department of Clinical
Research and Innovation, Assistance Publique–
Hôpitaux de Marseille, Marseille, France (Loubière,
Baumstarck, Loundou, Auquier); ENSEIS, Centre
Max Weber (UMR 5283), Lyon, France (Mougeot);
Laboratoire de recherche en Santé Mentale et
Sciences Humaines et Sociales (Labo SM-SHS), GHU
Paris Psychiatry Neurosciences, Paris, France
(Jouet); Resource Center of Psychosocial
Rehabilitation, Centre Hospitalier Le Vinatier, Lyon,
France (Franck); UMR 5229, Université de Lyon and
CNRS, Villeurbanne, France (Franck).
Author Contributions: Drs Tinland and Loubière
had full access to all of the data in the study and
take responsibility for the integrity of the data and
the accuracy of the data analysis.
Concept and design: Tinland, Loubière, Pontier,
Franck, Lançon, Auquier.
Acquisition, analysis, or interpretation of data:
Tinland, Loubière, Mougeot, Jouet, Baumstarck,
Loundou, Franck.
Drafting of the manuscript: Tinland, Loubière,
Mougeot, Franck.
Critical revision of the manuscript for important
intellectual content: Tinland, Loubière, Jouet,
Pontier, Baumstarck, Loundou, Franck, Lançon,
Auquier.
Statistical analysis: Tinland, Loubière, Loundou.
Obtained funding: Tinland, Loubière, Franck.
Administrative, technical, or material support:
Tinland, Jouet, Pontier, Franck.
Supervision: Tinland, Jouet, Franck, Lançon,
Auquier.
Conflict of Interest Disclosures: Drs Tinland,
Loubière, and Jouet reported grants from the
French Ministry of Health Directorate General of
Health Care Services (DGOS) during the conduct of
the study. No other disclosures were reported.
Funding/Support: This work is supported by an
institutional grant from the French 2017 National
Program of Health Services Research (Programme
de Recherche sur la Performance du système de
Soins, PREPS-17-0575, FINESS number
130786049). The Clinical Research Direction of
Assistance Publique Hôpitaux de Marseille is the
sponsor of this trial.
758 JAMA Psychiatry August 2022 Volume 79, Number 8 (Reprinted)
Downloaded From: https://jamanetwork.com/ on 06/12/2023
Effect of Peer Worker–Facilitated Psychiatric Advance Directives in People With Mental Illness
cant increase in some mental health outcomes (self-perceived
symptoms, empowerment, and recovery). These findings sup-
port the use of PW-PADs for people with schizophrenia, bipo-
lar I disorder, or schizoaffective disorder. Legal and organiza-
tional initiatives that promote supported decision-making can
develop the activity of peer workers to fulfill this mission.
Role of the Funder/Sponsor: The role of the
Clinical Research Direction of Assistance Publique
Hôpitaux de Marseille is to ensure legal and
administrative responsibility: guarantee compliance
with the legal framework, in particular the ethical
legal framework, transmit protocol amendments to
ethics committees, perform quality control for the
trial, report to the funder, and sign agreements with
hospitals. The list of study sites is available from this
service. The funding sources had no role in the
design or conduct of the study; collection, analysis,
or interpretation of the data; preparation, review, or
approval of the manuscript; or the decision to
submit the manuscript for publication.
Group Information: Additional members of the
DAiP Group appear in Supplement 3.
Meeting Presentation: This work was presented at
the European Congress of Psychiatry (EPA 2022);
June 6, 2022; Budapest, Hungary.
Data Sharing Statement: See Supplement 4.
Additional Contributions: We thank our funder,
the DGOS, which financed the entire study (PREPS
2017-0575), as well as our sponsor, Marseille Public
Hospital (APHM), and Emilie Garrido, director of
research, for her attention; Jean Dhorne for his
impeccable help in monitoring this clinical trial;
Marika Larose for continuing her monitoring as best
she could despite the pandemic; Mariola
Klimkowska for having facilitated the hiring; and
Richard Malkoun for the data management of the
study. We thank Celine Letailleur, EUTOPIA, the
MARSS team, and the CoFoR Recovery College for
their help and advice in developing the tool and
facilitating peer-worker practice sessions. We
warmly thank the peer workers, research assistants,
and intern of DAiP for their daily involvement:
Bastien Vincent, Iannis McCluskey, Camille Niard,
Oriane Beynel, Léa Leclerc, and Nicholas
Armstrong. We also thank Magali Coldefy and the
CONFCAP group, in particular Benoît Eyraud, and
also all the investigating psychiatrists: Jean Naudin,
PhD; Edouard Leaune, PhD; Sophie Cervello, MD;
Jacques Glikman, MD; Nathalie Christodoulou, MD;
Christophe Lamisse, MD; Pierre Giordano, MD; Yves
Guillermain, MD; Pierre Morcellet, MD; Emma
Beetlestone, MD; René Diouaba, MD; Marie
Degrandi, MD; Catherine Faget, MD; Raoul
Belzeaux, PhD; Michel Cermolacce, PhD; Sandrine
Limousin, MD; Marie Degrandi, MD; Laura
Brandejsky, MD; Julien Testard, MD; Marion Dubois,
MD; Xavier Zendjidjian, PhD; Lucie-Oriane Plazat,
MD; Héloïse Da Costa, MD; Maxence Bras, MD;
Christian Vedie, MD; Hélène Pigeon, MD; Nicole
Beer, MD; Gabrielle Durand, MD; Caroline Peiffer,
MD; Emma Wieviorka, MD; Mohamed Ali Ben
Mustapha, MD; Santiago Riascos-Henao, MD;
Isabelle Blondiaux, MD; Julie Meudal, MD; Alexis
Van Der Elst, MD; Olivia Vergely, MD; Dina Sakh,
MD; Annie Msellati, MD; Laurène Beherec, MD; and
Sophiane Chafaï, MD. English proofreading was
provided by American Journal Expert (AJE) and
Owen Taylor, and we thank them for their help.
Further, we are grateful to the participants.
Participants were compensated for meetings with
research assistants, and sponsor staff members,
peer workers, and research assistants were paid for
their contributions.
REFERENCES
1. Barber H. Seclusion, restraint and coercion:
abuse “far too common” in mental health services
across the world. The Telegraph. Published June 10,
2021. Accessed October 14, 2021. https://www.
telegraph.co.uk/global-health/climate-and-people/
seclusion-restraint-coercion-abuse-far-common-
mental-health/
2. World Health Organization. Guidance on
community mental health services: promoting
person-centred and rights-based approaches.
Accessed October 14, 2021. https://apps.who.int/
iris/handle/10665/341648
3. Rodrigues M, Hermann H, Galderisi S, Allan J.
Implementing alternatives to coercion: a key
component of improving mental health care. World
Psychiatric Association. Published October 2020.
Accessed November 1, 2021. https://3ba346de-
fde6-473f-b1da-536498661f9c.filesusr.com/ugd/
e172f3_635a89af889c471683c29fcd981db0aa.pdf
4. de Jong MH, Kamperman AM, Oorschot M, et al.
Interventions to reduce compulsory psychiatric
admissions: a systematic review and meta-analysis.
JAMA Psychiatry. 2016;73(7):657-664.
doi:10.1001/jamapsychiatry.2016.0501
5. Molyneaux E, Turner A, Candy B, Landau S,
Johnson S, Lloyd-Evans B. Crisis-planning
interventions for people with psychotic illness or
bipolar disorder: systematic review and
meta-analyses. BJPsych Open. 2019;5(4):e53.
doi:10.1192/bjo.2019.28
6. Swanson JW, Swartz MS, Elbogen EB, et al.
Facilitated psychiatric advance directives:
a randomized trial of an intervention to foster
advance treatment planning among persons with
severe mental illness. Am J Psychiatry. 2006;163
(11):1943-1951. doi:10.1176/ajp.2006.163.11.1943
7. Szmukler G. Treatment pressures, coercion and
compulsion in mental health care. J Ment Health.
2008;17(3):229-231. doi:10.1080/
09638230802156731
8. Hem MH, Gjerberg E, Husum TL, Pedersen R.
Ethical challenges when using coercion in mental
healthcare: a systematic literature review. Nurs Ethics.
2018;25(1):92-110. doi:10.1177/0969733016629770
9. Larsen IB, Terkelsen TB. Coercion in a locked
psychiatric ward: perspectives of patients and staff.
Nurs Ethics. 2014;21(4):426-436. doi:10.1177/
0969733013503601
10. Scheyett AM, Kim MM, Swanson JW, Swartz
MS. Psychiatric advance directives: a tool for
consumer empowerment and recovery. Psychiatr
Rehabil J. 2007;31(1):70-75. doi:10.2975/31.1.2007.
70.75
jamapsychiatry.com
 Effect of Peer Worker–Facilitated Psychiatric Advance Directives in People With Mental Illness
11. Papageorgiou A, King M, Janmohamed A,
Davidson O, Dawson J. Advance directives for
patients compulsorily admitted to hospital with
serious mental illness: randomised controlled trial.
Br J Psychiatry. 2002;181(6):513-519. doi:10.1192/
bjp.181.6.513
12. Ruchlewska A, Wierdsma AI, Kamperman AM,
et al. Effect of crisis plans on admissions and
emergency visits: a randomized controlled trial.
PLoS One. 2014;9(3):e91882. doi:10.1371/journal.
pone.0091882
13. Davidson L, Bellamy C, Guy K, Miller R. Peer
support among persons with severe mental
illnesses: a review of evidence and experience.
World Psychiatry. 2012;11(2):123-128. doi:10.1016/
j.wpsyc.2012.05.009
14. Easter MM, Swanson JW, Robertson AG, Moser
LL, Swartz MS. Facilitation of psychiatric advance
directives by peers and clinicians on Assertive
Community Treatment teams. Psychiatr Serv. 2017;
68(7):717-723. doi:10.1176/appi.ps.201600423
15. Belden CM, Gilbert AR, Easter MM, Swartz MS,
Swanson JW. Appropriateness of psychiatric
advance directives facilitated by peer support
specialists and clinicians on Assertive Community
Treatment teams. J Ment Health. 2022;31(2):239-245.
doi:10.1080/09638237.2021.1952946
16. Tinland A, Leclerc L, Loubière S, et al.
Psychiatric advance directives for people living with
schizophrenia, bipolar I disorders, or schizoaffective
disorders: study protocol for a randomized
controlled trial: DAiP study. BMC Psychiatry. 2019;
19(1):422. doi:10.1186/s12888-019-2416-9
17. American Psychiatric Association. Diagnostic
and Statistical Manual of Mental Disorders. 5th ed.
American Psychiatric Association; 2013.
18. Appelbaum PS, Grisso T. The MacArthur
Treatment Competence Study. I: Mental illness and
competence to consent to treatment. Law Hum
Behav. 1995;19(2):105-126. doi:10.1007/BF01499321
19. Misdrahi D, Verdoux H, Lançon C, Bayle F. The
4-Point ordinal Alliance Self-report: a self-report
questionnaire for assessing therapeutic
relationships in routine mental health. Compr
Psychiatry. 2009;50(2):181-185. doi:10.1016/
j.comppsych.2008.06.010
20. Auquier P, Simeoni MC, Sapin C, et al.
Development and validation of a patient-based
jamapsychiatry.com
Downloaded From: https://jamanetwork.com/ on 06/12/2023
health-related quality of life questionnaire in
schizophrenia: the S-QoL. Schizophr Res. 2003;63
(1-2):137-149. doi:10.1016/S0920-9964(02)
00355-9
21. EuroQol Group. EuroQol: a new facility for the
measurement of health-related quality of life.
Health Policy. 1990;16(3):199-208. doi:10.1016/
0168-8510(90)90421-9
22. Conrad KJ, Yagelka JR, Matters MD, Rich AR,
Williams V, Buchanan M. Reliability and validity of a
modified Colorado Symptom Index in a national
homeless sample. Ment Health Serv Res. 2001;3(3):
141-153. doi:10.1023/A:1011571531303
23. Rogers ES, Chamberlin J, Ellison ML, Crean T.
A consumer-constructed scale to measure
empowerment among users of mental health
services. Psychiatr Serv. 1997;48(8):1042-1047.
doi:10.1176/ps.48.8.1042
24. Labbé É, Moulin JJ, Guéguen R, Sass C, Chatain
C, Gerbaud L. Un indicateur de mesure de la
précarité et de la santé sociale: le score EPICES. La
Revue de l’Ires. 2007;53(1):3-49.
25. Thornicroft G, Farrelly S, Szmukler G, et al.
Clinical outcomes of Joint Crisis Plans to reduce
compulsory treatment for people with psychosis:
a randomised controlled trial. Lancet. 2013;381
(9878):1634-1641. doi:10.1016/S0140-6736(13)
60105-1
26. Henderson C, Flood C, Leese M, Thornicroft G,
Sutherby K, Szmukler G. Effect of joint crisis plans
on use of compulsory treatment in psychiatry:
single blind randomised controlled trial. BMJ. 2004;
329(7458):136. doi:10.1136/bmj.38155.585046.63
27. Plancke L, Amariei A, Flament C, Dumesnil C.
Psychiatric readmissions: individual and
organizational factors. Article in French. Sante
Publique. 2017;29(6):829-836. doi:10.3917/spub.176.
0829
28. van Buuren S. Multiple imputation of discrete
and continuous data by fully conditional
specification. Stat Methods Med Res. 2007;16(3):
219-242. doi:10.1177/0962280206074463
29. Schafer JL, Graham JW. Missing data: our view
of the state of the art. Psychol Methods. 2002;7
(2):147-177. doi:10.1037/1082-989X.7.2.147
30. Chieze M, Hurst S, Kaiser S, Sentissi O. Effects
of seclusion and restraint in adult psychiatry:
(Reprinted) JAMA Psychiatry August 2022 Volume 79, Number 8
Original Investigation Research
a systematic review. Front Psychiatry. 2019;10:491.
doi:10.3389/fpsyt.2019.00491
31. Tingleff EB, Bradley SK, Gildberg FA,
Munksgaard G, Hounsgaard L. “Treat me with
respect”: a systematic review and thematic analysis
of psychiatric patients’ reported perceptions of the
situations associated with the process of coercion.
J Psychiatr Ment Health Nurs. 2017;24(9-10):681-698.
doi:10.1111/jpm.12410
32. Nyttingnes O, Ruud T, Rugkåsa J. ‘It’s
unbelievably humiliating’: patients’ expressions of
negative effects of coercion in mental health care.
Int J Law Psychiatry. 2016;49(pt A):147-153.
doi:10.1016/j.ijlp.2016.08.009
33. Sheridan Rains L, Zenina T, Dias MC, et al.
Variations in patterns of involuntary hospitalisation
and in legal frameworks: an international
comparative study. Lancet Psychiatry. 2019;6(5):
403-417. doi:10.1016/S2215-0366(19)30090-2
34. Lay B, Kawohl W, Rössler W. Outcomes of a
psycho-education and monitoring programme to
prevent compulsory admission to psychiatric
inpatient care: a randomised controlled trial.
Psychol Med. 2018;48(5):849-860. doi:10.1017/
S0033291717002239
35. Barrenger SL, Stanhope V, Miller E. Capturing
the value of peer support: measuring
recovery-oriented services. J Public Ment Health.
2019;18(3):180-187. doi:10.1108/JPMH-02-2019-0022
36. Pitt V, Lowe D, Hill S, et al. Consumer-providers
of care for adult clients of statutory mental health
services. Cochrane Database Syst Rev. 2013;(3):
CD004807. doi:10.1002/14651858.CD004807.pub2
37. White S, Foster R, Marks J, et al. The
effectiveness of one-to-one peer support in mental
health services: a systematic review and
meta-analysis. BMC Psychiatry. 2020;20(1):534.
doi:10.1186/s12888-020-02923-3
38. Lloyd-Evans B, Mayo-Wilson E, Harrison B,
et al. A systematic review and meta-analysis of
randomised controlled trials of peer support for
people with severe mental illness. BMC Psychiatry.
2014;14:39. doi:10.1186/1471-244X-14-39
39. Bellamy C, Schmutte T, Davidson L. An update
on the growing evidence base for peer support.
Ment Health Soc Incl. 2017;21(3):161-167.
doi:10.1108/MHSI-03-2017-0014
759